134 ALASKAN GLACIER STUDIES
it and fell into moulins. The lower half of this distributary was stained with a thin film of de'bris, increasing in amount toward the end. The film formed a general sheet of clay and sand with some pebbles and bowlders, but was not thick enough to completely hide the ice. The finer material was being carried off in the streams. Here and there were pyramids of de'bris, with ice cores, varying in height from an inch to 6 or 8 feet, evidently on the site of earlier depressions in which the de'bris had gathered to later be etched into relief because of the protection against ablation which they gave to these parts of the wasting glacier surface. There were also numerous short ridges of similar character marking the sites of former crevasses. The outer edge of the glacier thinned down so that we could not determine its exact position. This edge was not covered by ablation moraine, but in places was buried beneath stratified deposits laid down by the water which escaped from the melting glacier. Some of these deposits were laid down in a temporary marginal lake, some as alluvial fan deposits; and all were laid down since 1899, for they occupied the site of the glacier end as photographed by Gilbert in that year.
A good sized stream emerged from the glacier front and flowed through the depression in which the stratified deposits lay, then entered a deep gorge in the friable slates and schists, which turns at right angles across the tail of the nunatak (Pig. 10; PL LXIII). Gilbert's photographs prove that in 1899 the glacier front extended into this gorge. After emerging from the gorge the stream spreads out in numerous branches over an alluvial fan which it is building in Nunatak Fiord, at the western pase of the nunatak. We have been puzzled to explain this gorge for, from Russell's description of the condition of Nunatak Glacier, it seems certain that, even though the two arms of the glacier may not have been actually united in 1891, the land arm must surely have covered at least the upper part of the gorge in order to give him the impression that the hill was a nunatak. The gorge is certainly too broad and deep for the work of such a glacial stream in the 14 years between 1891 and 1905. It may of course be the product of stream work before the last advance of the Nunatak Glacier, or it may in part have resulted from subglacial stream erosion when the land tongue extended down beyond the western end of the nunatak.
As compared with the condition in 1899 the land arm of Nunatak Glacier in 1905 was notably different in several respects. Both lateral moraine areas had broadened considerably, thus reducing the area of clear ice. At the same time, the extent of de'bris-stained ice between the lateral moraines had increased, evidently as a result of the lowering of the glacier surface, which was very noticeable by comparison with Gilbert's photographs. The glacier surface of 1905 was certainly as much as 100 feet lower than it was in 1899, and its front had receded not less than two or three hundred yards.
In 1909 we found evidence of the continuation of the same changes (PL LXI, B). There had not been a very great recession of the front, although there had been some, but the glacier surface was very much lower and the lateral moraine areas were still broader and more in relief than in 1905. The very outermost edge of the glacier was a low, flat sheet of debris-stained ice (PL LXII, A). This distributary is evidently a dying end, no longer in motion, and, unless an advance begins, its rate of recession and period of ultimate destruction will depend entirely on the rate at which ablation can remove the ice; and this rate, though not known quantitatively, is certainly very rapid. A few decades should destroy this tongue if no advance comes in the meantime.
There is evidence that it has been steadily receding for a long time. Far out beyond